Current highly active antiretroviral therapy (HAART) cannot eliminate HIV-1 from infected persons, mainly because of the existence of refractory viral reservoir(s). Beyond latently-infected CD4+-T lymphocytes, macrophages (M/M) are important persistent reservoirs for HIV in vivo, that represent a major obstacle to HIV-1 eradication. Therefore, a rational therapeutic approach directed to the selective elimination of long-living HIV-infected M/M may be relevant in the therapy of HIV infection. Here we report that HIV-1 chronic infection of human macrophages results in the marked increase of expression and phosphorylation of STAT1, a protein involved in the regulation of many functions such as cell growth, differentiation, and maintenance of cellular homeostasis, thereby providing a new molecular target for drug development. A single and brief exposure to 9-(beta-D-arabinofuranosyl)-2-fluoroadenine 5'-monophosphate (FaraAMP, Fludarabine), a potent antileukemic nucleoside analog active against STAT1 expressing cells, selectively kills macrophage cultures infected by HIV-1 without affecting uninfected macrophages. Furthermore, encapsulation of Fludarabine into autologous erythrocytes (RBC) and targeting to macrophages through a single-18 h treatment with drug-loaded RBC, not only abolishes the Fludarabine-mediated toxic effect on non-phagocytic cells, but also enhances the selective killing of HIV-infected macrophages. As a final result, a potent (>98%) and long-lasting (at least 4 weeks without rebound) inhibition of virus release from drug-loaded RBC-treated chronically-infected macrophages was achieved. Taken together, the evidence of HIV-1-induced increase of STAT1, and the availability of a selective drug targeting system, may prove useful in the design of new pharmacological treatments to clear the HIV-1 macrophage reservoir.